Mathematician predicts labs will abandon IMO-level math as AI makes it 'too easy'

burny_tech · x · 2026-09-06

Mathematician Alex Kontorovich recounts that AlphaProof came within one point of an IMO gold medal two summers ago, half a dozen labs claimed gold-level performance last summer — and he predicts zero systems will bother this year, because IMO has become 'too easy' for AI and nobody cares anymore. He extends the prediction: next summer labs will solve zero major math problems, returning them to professors and grad students. The reposter adds that interest will shift toward applied mathematics — using AI to model real-world problems with constraints — while compute remains the real bottleneck.

Related event: Mathematician Predicts AI IMO Gold Will Soon Be Ignored(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →